The sync job

The sync job is an action that takes place repeatedly throughout the life-time of the replica. When the sync job takes place, it creates a snapshot on the source dataset and delivers it to the target. The next sync job takes a new snapshot of the source, calculates the diff from the previous snapshot and sends only the data that was changed since the previous sync job. The replica creates time stamps on a dataset level.  

The amount of time between two planned sync jobs is called sync interval. The sync interval can be changed by the user. Changes to the replica sync interval take effect on the next sync job. 

The replica states

InfiniBox uses replica states in order to manage a prompt functionality of the replication. The replica starts with an initialization state, where the source volume is copied to the target in order to generate a consistent snapshot. Once the Initialization ends, the target is consistent

Please note that when a large number of replicas are initialized at the same time, the system processes them – initializing the replication – one by one. Some of the replica states will not be changed for a while.

The replica moves between an Idle and Replicating states. On Replicating, a sync job is running. On Idle, the replica is waiting for the next sync job.

Whenever there is a network failure, the replica will remain in Replicating state and continue once the network connection is back. If the network failure lasts more than 24 hours, the replica enters to an Auto Suspend state.

Auto Suspend can also occur as a result of a permanent failure. An example for a permanent failure is pool space depletion (see below). The replica exits the Auto Suspend state by manually issuing a resume command.

In addition, there is an option to manually suspend the replica. The manual suspension of the replica stops the current sync job and cancels additional sync jobs.

Pool space depletion

Throughout the replica life-cycle, the capacity of the pool - either on the source or the target - can be depleted. In such a case, the replication is auto-suspended and manual intervention is required.

This situation is handled by either expanding the pool, or moving the volume to another pool, making room for new data to be written and replicated. 

Permissions to provision link and replica

The ability to manage replicas and links is based on the permission level of the user that performs the operation. The user has to have permissions to both source and target systems, in order to:

  • Link between the systems
  • Create a replica and other replica operations

When trying to perform an action on the remote system, the system tries to connect to the remote system with the same credentials as the local. If the credentials are not identical on both systems, the user has to specify different credentials for the remote system.

Rate limiting on the physical interface

InfiniBox allows capping the throughput on the physical interface in order to refrain from overcrowding the line between the system sites.

Support for INFINIDAT Storage Replication Adapter

The INFINIDAT Storage Replication Adapter is a software add-on that integrates with the VMware vCenter Site Recovery Manager (SRM) platform, enabling site-to-site failovers over InfiniBox systems that are deployed on remote sites.

For instructions on how to use the INFINIDAT Storage Replication Adapter, see: InfiniBox Storage Replication Adapter.

Advanced support for disaster recovery test scenarios

The target volume is write-protected as long as the replica is active. For DR tests scenarios, InfiniBox supports taking snapshots of the target volume. The snapshot that is taken on the target volume is always consistent with the last replication cycle.

A snapshot can be taken on the target volume while there is a sync job running at the same time. In such a case, the snapshot will be consistent with the previous sync job.

This ability to take a snapshot on the target is ideal for Disaster Recovery tests that aim at verifying the integrity of the data on the target without affecting the replication process.






Was this article helpful?
0 out of 0 found this helpful

0 out of 0 found this helpful